Re:I Used Stable Diffusion and Dreambooth to Create an Art P...
by AuryGlenz on Sunday April 16, 2023 @03:42PM
(#35593601)
In my (limited) experience, dogs seem to be easier than people for fine-tuning - especially if your end result is going to be artsy. Faces of people you know well being off in slight ways really throws you off, but with dogs there's a bit more leeway.
It wasn't actually particularly hard - I used a Colab notebook on the free tier to fine-tune the model, and even got chatGPT to write some of the prompts.